The development of software packages designed to permit simplified graphical user interface (GUI)-based access to the wealth of electronic information available over the Internet and the World Wide Web has lead to a dramatic increase in the amount of information that is currently available over public computer networks. Unlike the highly controlled atmosphere of a private computer network, however, it is difficult to monitor user interaction with network resources on public networks. As a result, it is difficult for individual servers on a public network to determine how long individual users have interacted with their resources, or how much information has been downloaded. It is equally difficult for individual servers to target specialized information to a particular audience or to learn the identity of individual users on a public network.
The techniques utilized in many private networks for monitoring client use and interaction do not lend themselves to public networks. For example, user access to a server in private networks is generally obtained through the use of a unique identification number provided by the server. Details of individual user interaction with the network are closely monitored by server-resident processes, and historic databases are automatically generated and continually updated to track the nature and amount of information accessed by individual users, as well as their connection time. This information is generally used, for example, to maintain a subscriber-indexed billing database.
In a public computer network, however, use of server-resident monitoring techniques may be severely limited. In some public networks, subscribers are given unlimited access, via a service provider, to a virtually unlimited number of servers, and no permanent connection is usually made between these servers and a client machine. The nature and amount of information downloaded by individual users is not easily monitored for each client machine and only limited information concerning individual user interaction with the network may generally be captured by a server (i.e., so-called network ID and client ID).
Due largely to the lack of advanced monitoring techniques available to individual servers on a public network, the same information is generally served out to all clients on a completely untargeted basis. In other words, the same information is generally downloaded to all users that access a particular resource on a server, irrespective of individual user interests. There is therefore a need to provide servers on a public network with the ability to automatically monitor use of and interaction with resources downloaded by users so as to facilitate the targeted serving of information.
While various methods are known for obtaining information concerning user preferences, no such methods are automatic. For instance, one application, known as a “customizable home page”, permits users, upon the request of a server, to make certain choices. When a user who has done so contacts that server at a later date, the server assembles information for downloading to the user in accordance with the previously-selected choices. More specifically, the user visits a so-called “Web page” of a particular server where he or she is asked to fill in a blank form by selecting various preferences, such as links to favorite Web sites, interests in entertainment, sports, and the like. The user then submits this information to the server by clicking the so-called “submit” button of the fill-in form, which causes the client to transmit the information to the server. The server returns a Web page with a response header which creates, or “sets” an ID field located in a file on the client computer (this file is known as the “client ID” or “cookie”) to include information about the user's preferences. When the user later returns to a specified Uniform Resource Locator, or “URL”, on the same server, the “client ID” or “cookie” with the previously-set preference information is transmitted in the HTTP request header to the server, which can then return a Web page that is assembled according to the user-specific information. This application is disclosed, for example, in A. Gundavaram, CGI Programming on the World Wide Web, O'Reilly Press, 1996.
While the “customizable home page” facilitates the serving of information on a limited targeted basis, it does not provide for the automatic determination of user interests, and inconveniences the user by requesting that he or she specify various preferences. Moreover, use of a customizable home page is limited to individual Web sites and can not be “spread out” over multiple resources on different servers. In other words, while a customizable home page may be of use with respect to the particular resources located on a single server, it does not serve any purpose for other servers on a public network. A variation of this technique is used by some servers to download executable programs. For instance, one such application disclosed by G. Cornell and C. S. Horstmann, in Core Java, The SunSoft Press, 1996, involves the generation of “order forms” on client computers. In this application, the client machine loads a Web page from a server which has an embedded link to an executable program that downloads to and executes on the client machine. Upon execution in the client machine, the program contacts the server and retrieves a list of goods and associated prices. The program allows the user to order various goods and requires the user to fill out a form for billing purposes. The user “clicks” on the submit button of the fill-in form to transmit the information to the server. Like the customizable home page, this method of user-specific data acquisition requires the active participation of the user, and does not provide for the automatic determination of user preferences and interests.
In addition to the inability to serve out information on a targeted basis, which is of enormous concern from a marketing standpoint, the limited monitoring capabilities available to individual servers makes it difficult for servers and administrators to determine how long users have viewed their resources and how much information has been downloaded by individual users so as to be able to bill client use and interaction with network resources and to analyze the value and effectiveness of such resources. As a result, much of the information provided by a server over a public network is the same for all clients. In addition, while it is currently possible to track a user's links within the same resource, there is no standard way to track user's links across multiple resources on different servers. For example, a common occurrence in public networks is when a user is viewing a first resource and “clicks on” a link to a second resource located on a different server. In such instances, the second resource is downloaded and the first resource is either discarded or held in background. However, there is generally no uniform way in which to monitor such occurrences. In addition, while it is currently possible to track the number of times a particular resource has been accessed, it has generally not been possible to track the length of time a particular resource has been viewed by a particular user. There is also a great deal of other valuable information concerning user interaction with a resource which would be useful to administrators, advertisers, marketing professionals and the like, but which can not be conveniently collected using current monitoring techniques.
For example, one of the largest public networks, the “Internet”, has become an extremely popular advertising tool. Many companies have their own Internet “Web sites” and have also purchased advertising space within more popular Web sites of other companies. For instance, many advertisers purchase so-called “advertising banner” (or “ad banner”) space within the Web page of a popular site, thereby allowing consumers to “click-through” (i.e., specify a link) to the Web site of the advertiser. In many cases, the use of an ad banner substantially increases the advertiser's exposure. Using the limited monitoring techniques available to Internet servers, however, it is difficult to determine the effectiveness of individual Web sites and ad banners. For instance, known monitoring techniques are generally limited to determining the number of times a Web page was downloaded. Similar techniques are used to determine the number of times an ad banner (which is embedded inside a Web page) has been displayed, and how many times the banner was “clicked” on to visit the Web site of the advertiser.
Generally, an ad banner is embedded inside a Web page located on a first server through the use of the known HTML <IMG> tag. When a client machine passes a TCP/IP request for the Web page to the first server, the Web page is downloaded to the client, including the ad banner embedded using the <IMG> tag. The <IMG> tag is used to reference a resource (i.e., the “ad banner”) stored on the same or a different server which captures the user's ID (via the HTTP request header) and dynamically returns an ad related image to the client for display within the Web page. At the same time, a counter representing the number of times the specific ad has been displayed is incremented. The ad banner itself may have an embedded address referring to yet another Web resource. In such an instance, if the user “clicks” on the ad banner, the client may load a resource on the second server which once again captures the user's ID and forwards the user to a Web resource which is appropriate for the displayed ad (for example, a page on the advertiser's Web site). At the same time, a counter representing the number of times the specific ad was clicked on to go to the advertiser's Web site is incremented.
While Web sites and ad banners have, in some cases, been valuable marketing tools, the limited monitoring capabilities available to servers on networks in which no permanent connection is made between a server and a client (such as the Internet) has prevented these marketing tools from being used to their full potential. Since HTTP or Web servers cannot automatically determine the amount of time and the frequency at which particular users interact with their resources, Web site administrators and advertisers cannot accurately determine the effectiveness of their resources. Since servers cannot automatically monitor user interaction and automatically obtain user preferences and interests, servers cannot assemble and serve resources targeted to individual user interests.